DCGAN is initialized with random weights, so a random code plugged into your network would create a totally random graphic. However, while you may think, the network has countless parameters that we could tweak, and also the purpose is to locate a location of such parameters that makes samples created from random codes seem like the schooling detai